Examples with FileSystem edu.iu.dsc.tws.api.data.FileSystem used on opensource projects

Example 21 with FileSystem

use of edu.iu.dsc.tws.api.data.FileSystem in project twister2 by DSC-SPIDAL.

the class FileInputPartitioner method createInputSplits.

/**
 * Computes the input splits for the file. By default, one file block is one split. If more splits
 * are requested than blocks are available, then a split may be a fraction of a
 * block and splits may cross block boundaries.
 *
 * @param minNumSplits The minimum desired number of file splits.
 * @return The computed file splits.
 */
@Override
public FileInputSplit<OT>[] createInputSplits(int minNumSplits) throws IOException {
    if (minNumSplits < 1) {
        throw new IllegalArgumentException("Number of input splits has to be at least 1.");
    }
    // take the desired number of splits into account
    int curminNumSplits = Math.max(minNumSplits, this.numSplits);
    final Path path = this.filePath;
    final List<FileInputSplit> inputSplits = new ArrayList<FileInputSplit>(curminNumSplits);
    // get all the files that are involved in the splits
    List<FileStatus> files = new ArrayList<FileStatus>();
    long totalLength = 0;
    // final FileSystem fs = path.getFileSystem();
    final FileSystem fs = FileSystemUtils.get(path, config);
    final FileStatus pathFile = fs.getFileStatus(path);
    if (pathFile.isDir()) {
        totalLength += sumFilesInDir(path, files, true);
    } else {
        // TODO L3: implement test for unsplittable
        // testForUnsplittable(pathFile);
        files.add(pathFile);
        totalLength += pathFile.getLen();
    }
    // TODO L3: Handle if unsplittable
    // TODO L1: check if we can add the i j method when making splits so that the last split is not
    // larger than the other splits
    final long maxSplitSize = totalLength / curminNumSplits + (totalLength % curminNumSplits == 0 ? 0 : 1);
    // Generate the splits
    int splitNum = 0;
    for (final FileStatus file : files) {
        final long len = file.getLen();
        final long blockSize = file.getBlockSize();
        final long localminSplitSize;
        if (this.minSplitSize <= blockSize) {
            localminSplitSize = this.minSplitSize;
        } else {
            LOG.log(Level.WARNING, "Minimal split size of " + this.minSplitSize + " is larger than the block size of " + blockSize + ". Decreasing minimal split size to block size.");
            localminSplitSize = blockSize;
        }
        final long splitSize = Math.max(localminSplitSize, Math.min(maxSplitSize, blockSize));
        final long halfSplit = splitSize >>> 1;
        final long maxBytesForLastSplit = (long) (splitSize * MAX_SPLIT_SIZE_DISCREPANCY);
        if (len > 0) {
            // get the block locations and make sure they are in order with respect to their offset
            final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, len);
            Arrays.sort(blocks);
            long bytesUnassigned = len;
            long position = 0;
            int blockIndex = 0;
            while (bytesUnassigned > maxBytesForLastSplit) {
                // get the block containing the majority of the data
                blockIndex = getBlockIndexForPosition(blocks, position, halfSplit, blockIndex);
                // create a new split
                FileInputSplit fis = createSplit(splitNum++, file.getPath(), position, splitSize, blocks[blockIndex].getHosts());
                inputSplits.add(fis);
                // adjust the positions
                position += splitSize;
                bytesUnassigned -= splitSize;
            }
            if (bytesUnassigned > 0) {
                blockIndex = getBlockIndexForPosition(blocks, position, halfSplit, blockIndex);
                final FileInputSplit fis = createSplit(splitNum++, file.getPath(), position, bytesUnassigned, blocks[blockIndex].getHosts());
                inputSplits.add(fis);
            }
        } else {
            // special case with a file of zero bytes size
            final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, 0);
            String[] hosts;
            if (blocks.length > 0) {
                hosts = blocks[0].getHosts();
            } else {
                hosts = new String[0];
            }
            final FileInputSplit fis = createSplit(splitNum++, file.getPath(), 0, 0, hosts);
            inputSplits.add(fis);
        }
    }
    LOG.fine("input splits value:" + inputSplits.size() + "\t" + Arrays.toString(inputSplits.toArray()));
    return inputSplits.toArray(new FileInputSplit[inputSplits.size()]);
}

Example 22 with FileSystem

use of edu.iu.dsc.tws.api.data.FileSystem in project twister2 by DSC-SPIDAL.

the class CompleteArrowInputPartitioner method createInputSplits.

/**
 * It creates the split for the complete file.
 *
 * @param minNumSplits Number of minimal input splits, as a hint.
 */
@Override
public FileInputSplit<OT>[] createInputSplits(int minNumSplits) throws IOException {
    if (minNumSplits < 1) {
        throw new IllegalArgumentException("Number of input splits has to be at least 1.");
    }
    int curminNumSplits = Math.max(minNumSplits, this.numSplits);
    final Path path = this.filePath;
    final List<FileInputSplit> inputSplits = new ArrayList<>(curminNumSplits);
    List<FileStatus> files = new ArrayList<>();
    long totalLength = 0;
    final FileSystem fs = FileSystemUtils.get(path, config);
    final FileStatus pathFile = fs.getFileStatus(path);
    if (pathFile.isDir()) {
        totalLength += sumFilesInDir(path, files, true);
    } else {
        files.add(pathFile);
        totalLength += pathFile.getLen();
    }
    final long maxSplitSize = totalLength;
    // Generate the splits
    int splitNum = 0;
    for (final FileStatus file : files) {
        final long len = file.getLen();
        final long blockSize = file.getBlockSize();
        final long localminSplitSize;
        if (this.minSplitSize <= blockSize) {
            localminSplitSize = this.minSplitSize;
        } else {
            LOG.log(Level.WARNING, "Minimal split size of " + this.minSplitSize + " is larger than the block size of " + blockSize + ". Decreasing minimal split size to block size.");
            localminSplitSize = blockSize;
        }
        final long splitSize = Math.max(localminSplitSize, Math.min(maxSplitSize, blockSize));
        if (len > 0) {
            final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, len);
            Arrays.sort(blocks);
            long position = 0;
            int blockIndex = 0;
            for (int i = 0; i < curminNumSplits; i++) {
                blockIndex = getBlockIndexForPosition(blocks, position, splitSize, blockIndex);
                final FileInputSplit fis = new CSVInputSplit(splitNum++, file.getPath(), position, splitSize, blocks[blockIndex].getHosts());
                inputSplits.add(fis);
            }
        } else {
            // special case with a file of zero bytes size
            final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, 0);
            String[] hosts;
            if (blocks.length > 0) {
                hosts = blocks[0].getHosts();
            } else {
                hosts = new String[0];
            }
            for (int i = 0; i < curminNumSplits; i++) {
                final FileInputSplit fis = new CSVInputSplit(splitNum++, file.getPath(), 0, 0, hosts);
                inputSplits.add(fis);
            }
        }
    }
    return inputSplits.toArray(new FileInputSplit[inputSplits.size()]);
}

Example 23 with FileSystem

use of edu.iu.dsc.tws.api.data.FileSystem in project twister2 by DSC-SPIDAL.

the class CompleteArrowInputPartitioner method sumFilesInDir.

/**
 * Enumerate all files in the directory and recursive if enumerateNestedFiles is true.
 *
 * @return the total length of accepted files.
 */
long sumFilesInDir(Path path, List<FileStatus> files, boolean logExcludedFiles) throws IOException {
    final FileSystem fs = FileSystemUtils.get(path);
    long length = 0;
    for (FileStatus file : fs.listFiles(path)) {
        if (file.isDir()) {
            if (acceptFile(file) && enumerateNestedFiles) {
                length += sumFilesInDir(file.getPath(), files, logExcludedFiles);
            } else {
                if (logExcludedFiles) {
                    LOG.log(Level.INFO, "Directory " + file.getPath().toString() + " did not pass the " + "file-filter and is excluded.");
                }
            }
        } else {
            if (acceptFile(file)) {
                files.add(file);
                length += file.getLen();
            } else {
                if (logExcludedFiles) {
                    LOG.log(Level.INFO, "Directory " + file.getPath().toString() + " did not pass the file-filter and is excluded.");
                }
            }
        }
    }
    return length;
}

Example 24 with FileSystem

use of edu.iu.dsc.tws.api.data.FileSystem in project twister2 by DSC-SPIDAL.

the class CompleteFileInputPartitioner method createInputSplits.

/**
 * It creates the split for the complete file.
 *
 * @param minNumSplits Number of minimal input splits, as a hint.
 */
@Override
public FileInputSplit<OT>[] createInputSplits(int minNumSplits) throws IOException {
    if (minNumSplits < 1) {
        throw new IllegalArgumentException("Number of input splits has to be at least 1.");
    }
    int curminNumSplits = Math.max(minNumSplits, this.numSplits);
    final Path path = this.filePath;
    final List<FileInputSplit> inputSplits = new ArrayList<>(curminNumSplits);
    List<FileStatus> files = new ArrayList<>();
    long totalLength = 0;
    final FileSystem fs = FileSystemUtils.get(path, config);
    final FileStatus pathFile = fs.getFileStatus(path);
    if (pathFile.isDir()) {
        totalLength += sumFilesInDir(path, files, true);
    } else {
        files.add(pathFile);
        totalLength += pathFile.getLen();
    }
    final long maxSplitSize = totalLength;
    // Generate the splits
    int splitNum = 0;
    for (final FileStatus file : files) {
        final long len = file.getLen();
        final long blockSize = file.getBlockSize();
        final long localminSplitSize;
        if (this.minSplitSize <= blockSize) {
            localminSplitSize = this.minSplitSize;
        } else {
            LOG.log(Level.WARNING, "Minimal split size of " + this.minSplitSize + " is larger than the block size of " + blockSize + ". Decreasing minimal split size to block size.");
            localminSplitSize = blockSize;
        }
        final long splitSize = Math.max(localminSplitSize, Math.min(maxSplitSize, blockSize));
        if (len > 0) {
            final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, len);
            Arrays.sort(blocks);
            long position = 0;
            int blockIndex = 0;
            for (int i = 0; i < curminNumSplits; i++) {
                blockIndex = getBlockIndexForPosition(blocks, position, splitSize, blockIndex);
                FileInputSplit fis = createSplit(splitNum++, file.getPath(), position, splitSize, blocks[blockIndex].getHosts());
                inputSplits.add(fis);
            }
        } else {
            // special case with a file of zero bytes size
            final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, 0);
            String[] hosts;
            if (blocks.length > 0) {
                hosts = blocks[0].getHosts();
            } else {
                hosts = new String[0];
            }
            for (int i = 0; i < curminNumSplits; i++) {
                final FileInputSplit fis = createSplit(splitNum++, file.getPath(), 0, 0, hosts);
                inputSplits.add(fis);
            }
        }
    }
    return inputSplits.toArray(new FileInputSplit[inputSplits.size()]);
}

Example 25 with FileSystem

use of edu.iu.dsc.tws.api.data.FileSystem in project twister2 by DSC-SPIDAL.

the class FixedInputPartitioner method sumFilesInDir.

long sumFilesInDir(Path path, List<FileStatus> files, boolean logExcludedFiles) throws IOException {
    final FileSystem fs = FileSystemUtils.get(path);
    long length = 0;
    for (FileStatus file : fs.listFiles(path)) {
        if (file.isDir()) {
            if (acceptFile(file) && enumerateNestedFiles) {
                length += sumFilesInDir(file.getPath(), files, logExcludedFiles);
            } else {
                if (logExcludedFiles) {
                    LOG.log(Level.INFO, "Directory " + file.getPath().toString() + " did not pass the " + "file-filter and is excluded.");
                }
            }
        } else {
            if (acceptFile(file)) {
                files.add(file);
                length += file.getLen();
            } else {
                if (logExcludedFiles) {
                    LOG.log(Level.INFO, "Directory " + file.getPath().toString() + " did not pass the file-filter and is excluded.");
                }
            }
        }
    }
    return length;
}